BOUWSMA Beery
2004-Aug-06 14:57 UTC
[icecast-dev] src/net/resolver.c patches for better IPv6 resolution
[caution: this is an IPv6 e-mail address. if you don't have IPv6-capable mail then best keep replies on the list only]> >existing getipnodebyname() call fails with IPv4 on my machine)... > Well, if you can make it work properly - send your patches in, please.The following enhancements to src/net/resolver.c (which looks identical across icecast, ices, libshout, and who knows what all) suffice here to get the test_resolver.c program to produce useful output on my FreeBSD machine -- if it's true that (some) Linux doesn't have getipnodebyname() then a different solution might be needed. The results when built with -DHAVE_IPV6=1 -DHAVE_GETIPNODEBYNAME=1 added to the Makefile DEFS on my FreeBSD 4.x: I got 172.27.72.27, when looking up stable.netscum.dyndns.dk. I got 2002:50da:100d:0:250:daff:fe21:edca, when looking up ipv6.netscum.dyndns.d k. I got 2002:50da:100d:0:250:daff:fe21:edca, when looking up netscum.dyndns.dk. I got ::1, when looking up localhost. I got news-feed02.inet.tele.dk, when looking up 193.162.153.122. I got swi6netCE1-A2-0-1.switch.ch, when looking up 2001:620:0:39::1. I got localhost, when looking up ::1. Much more satisfying than before. The results without the above added DEFS: I got (null), when looking up ipv6.netscum.dyndns.dk. I got 80.218.16.13, when looking up netscum.dyndns.dk. I got news-feed02.inet.tele.dk, when looking up 193.162.153.122. I got (null), when looking up 10.0.2.53. I got 2001:620:0:39::1, when looking up 2001:620:0:39::1. Now the question, for which I started to hack a bit: On an IPv6-aware system, _isip("2001:620:0:39::1") is true. On a non-IPv6-aware system, it's false, even though that's an IP of the IPv6 family. Should it be true? And if it's false, as you can see, it's interpreted as a hostname, and returned by resolver_getname(), instead of returning (null). Should I bother with this at all? (For example, should resolver_getname() verify that what it's returning is in fact a hostname and not garbage?) <p>Anyway, the patch. I'm sure there's a better way to do this, as I don't know what I'm doing, so I'd like to know what IPv6-capable OSen aren't able to get proper results -- it appears to work on my FreeBSD machine... Did I say I have no clue what I'm doing yet? <p>--- resolver.c-ORIG Mon Aug 5 16:48:03 2002 +++ resolver.c Sun Sep 29 19:43:54 2002 @@ -42,6 +42,13 @@ static mutex_t _resolver_mutex; static int _initialized = 0; +/* ugly hacks */ +struct in_addr inp; +union { + struct in_addr v4addr; + struct in6_addr v6addr; +} addr_u; + char *resolver_getname(const char *ip, char *buff, int len) { if (!_isip(ip)) { @@ -65,18 +72,11 @@ static int _isip(const char *what) { #ifdef HAVE_IPV6 - union { - struct in_addr v4addr; - struct in6_addr v6addr; - } addr_u; - if (inet_pton(AF_INET, what, &addr_u.v4addr) <= 0) return inet_pton(AF_INET6, what, &addr_u.v6addr) > 0 ? 1 : 0; return 1; #else - struct in_addr inp; - return inet_aton(what, &inp); #endif } @@ -87,27 +87,56 @@ #ifdef HAVE_GETIPNODEBYNAME int err; #else - struct in_addr inp; + char *temp; #endif struct hostent *host = NULL; - char *temp; /* do a little sanity checking */ if (what == NULL || buff == NULL || len <= 0) return NULL; #ifdef HAVE_GETIPNODEBYNAME - host = getipnodebyname(what, AF_INET6, AI_DEFAULT, &err); - if (host) { - if (_isip(what)) + if (_isip(what)) { + /* According to FreeBSD's man page, getipnodebyname and such + * aren't threadsafe either */ + _lock_resolver(); + host = getipnodebyaddr(&addr_u.v4addr, sizeof(struct in_addr), + AF_INET, &err); + _unlock_resolver(); + if (host == NULL) { + _lock_resolver(); + host = getipnodebyaddr(&addr_u.v6addr, + sizeof(struct in6_addr), AF_INET6, &err); + _unlock_resolver(); + } + if (host == NULL) { + buff = NULL; + } else { strncpy(buff, host->h_name, len); - else - inet_ntop(host->h_addrtype, host->h_addr_list[0], buff, len); - - freehostent(host); - } else - buff = NULL; -#else + freehostent(host); + } + } else { + /* According to FreeBSD's man page, getipnodebyname and such + * aren't threadsafe either */ + _lock_resolver(); + /* Try IPv6 if available and configured; else try IPv4 */ + host = getipnodebyname(what, AF_INET6, AI_ADDRCONFIG, &err); + _unlock_resolver(); + if (host == NULL) { + _lock_resolver(); + host = getipnodebyname(what, AF_INET, AI_ADDRCONFIG, + &err); + _unlock_resolver(); + } + if (host) { + inet_ntop(host->h_addrtype, host->h_addr_list[0], + buff, len); + freehostent(host); + } else { + buff = NULL; + } + } +#else /* HAVE_GETIPNODEBYNAME */ if (_isip(what)) { /* gotta lock calls for now, since gethostbyname and such * aren't threadsafe */ @@ -133,7 +162,7 @@ } } -#endif +#endif /* HAVE_GETIPNODEBYNAME */ return buff; } thanks, barry bouwsma, beer streamer --- >8 ---- List archives: http://www.xiph.org/archives/ icecast project homepage: http://www.icecast.org/ To unsubscribe from this list, send a message to 'icecast-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Hello all, When I want to retrieve the stats.xml file, the browser sends me an authenticxation popup. Perfect. The only thing, I can't find the location where username for this login is defined and set. Browsing through the src code, I tracked it down to the config.h file. It seems for authentication a config variable 'ice_login' is used. However, it is not mentioned in the config file icecast.xml. I tried adding it myself, but this does not work. Is this a bug or something under construction? The authentication to config.h is added in revision 1.8, but it seems to me the icecast.xml is not updated to this. ...or am I simply overlooking some things? Regards, Frank --- >8 ---- List archives: http://www.xiph.org/archives/ icecast project homepage: http://www.icecast.org/ To unsubscribe from this list, send a message to 'icecast-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Hi: The authentication for the xml stats is as follows: username: source password: <the password you've put in the config file> The icelogin item is to allow older style source logins (i.e. not the http authentication that was implemented a number of weeks ago). this login is deprecated, so it's been deliberately left out of the XML configuration example. Geoff. <p> -- Geoff Shang <gshang@uq.net.au> ICQ number 43634701 Make sure your E-mail can be read by everyone! http://www.betips.net/etc/evilmail.html Please avoid sending me Word or PowerPoint attachments. See http://www.fsf.org/philosophy/no-word-attachments.html <p>--- >8 ---- List archives: http://www.xiph.org/archives/ icecast project homepage: http://www.icecast.org/ To unsubscribe from this list, send a message to 'icecast-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
At 12:33 AM 10/11/02 +0200, you wrote:> >Hello all, > >When I want to retrieve the stats.xml file, the browser sends me an >authenticxation popup. Perfect. > >The only thing, I can't find the location where username for this login is >defined and set. Browsing through the src code, I tracked it down to the >config.h file.Currently, "source". That'll be configurable later. The ice_login stuff isn't really relevant (it's also deprecated). Mike --- >8 ---- List archives: http://www.xiph.org/archives/ icecast project homepage: http://www.icecast.org/ To unsubscribe from this list, send a message to 'icecast-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Apparently Analagous Threads
- IPv6 support for Icecast server
- [SAMBA-SECURITY] CVE-2007-0453: Buffer overrun in nss_winbind.so.1 on Solaris
- [SAMBA-SECURITY] CVE-2007-0453: Buffer overrun in nss_winbind.so.1 on Solaris
- ice2 CVS build problems under Solaris 7
- openssh-3.4p1 install problems on Solaris 6 with openssl-0.9.6g